Defense against pathogens and parasites requires substantial investment of energy and resources on part of the host. This makes the host immune function dependent on availability and accessibility of resources. A resource deprived host is therefore expected to be more susceptible to infections, although empirical results do not always align with this prediction. Limiting host access to resources can additionally impact within-host pathogen numbers, either directly by altering the amount of resources available to the pathogens for proliferation or indirectly by altering the efficiency of the host immune system. We tested for the effects of host starvation (complete deprivation of resources) on susceptibility to bacterial pathogens, and within-host pathogen proliferation, in Drosophila melanogaster females. Our results show that starvation increases post-infection mortality of the host, but in a pathogen-specific manner. This increase in mortality is always accompanied by increased within-host pathogen proliferation. We therefore propose that starvation compromises host resistance to bacterial infections in Drosophila melanogaster females thereby increasing susceptibility to infections.

1. Fly populations and general handling

Experiments reported here were carried out on flies from a large, outbred laboratory adapted population of Drosophila melanogaster, LH (Chippindale and Rice 2001, Prasad et al., 2007, Nandy et al., 2012). The LH population is maintained on a 14-day discrete generation cycle, at 25 ^OC temperature and 12:12 hour light-dark cycle, on cornmeal-molasses-yeast medium, at a census size of about 1900 adults. The flies are maintained in vials (95 mm height and 25 mm diameter); each generation starts with setting up of 60 vials with 150 eggs each on 8-10 ml of food medium. 12 days post-egg laying (PEL), by which time most adults have eclosed, adults from different vials are mixed together and redistributed into 60 vials with 16 females and 16 males in each vial. The vials are supplied with limiting live dietary yeast supplement, and on 14^th day PEL, the adults are transferred to fresh vials and allowed to oviposit for 18 hours to start the next generation. Vials with eggs greater than 150 undergo egg-culling to maintain the specified egg density in vials and avoid crowding.

2. Derivation of experimental flies

12-day PEL adults were transferred to plexiglass cages (14 cm length × 16 cm width × 13 cm height) at a density of 1000-1200 flies, and the cages are provided with standard food medium in 60 mm Petri plates. For collection of eggs for setting up experiments, cages are provided with a fresh food plate, supplemented with ad libitum live yeast supplement, for 48 hours. This is done to encourage egg production and laying in the females. Following this, a fresh food plate is provided to the cages and 12-14 hours later eggs are collected off these plates (using moist paint brushes on 1.5% agar gel) and seeded into food vials (with 8-10 ml of food medium) at an exact density of 150 eggs per vial. The number of vials set up in this manner depends upon the requirement for a particular experiment. These vials are then incubated under standard conditions (detailed above) for egg to mature into larvae and then into adults. On 10^th day PEL, during the eclosion peak, adults are collected in unmated state within 5-6 hours of eclosion, and housed in single-sex vials (each with 1-2 ml of food medium) at constant density of 8 females per vial or 10 males per vial. Flies are housed in these vials till further manipulation/experimentation.

3. Bacterial handling and infection protocol

The bacterial isolates are preserved as glycerol stocks at -80 ^OC. To obtain live bacterial cells for infections, 10 ml lysogeny broth (Luria Bertani Broth, Miler, HiMedia) is inoculated with glycerol stocks of the necessary bacterium, and incubated overnight with aeration (150 rpm shaker incubator) at suitable temperature. 100 microliters from this primary culture is inoculated into 10 ml fresh lysogeny broth and incubated for the necessary amount of time to obtain confluent (OD₆₀₀ = 1.0-1.2) cultures. The bacterial cells are pelleted down using centrifugation and resuspended in sterile MgSO₄ (10 mM) buffer at optical density (OD₆₀₀) of 1.0. Flies are infected, under light CO₂ anaesthesia, by pricking them on the dorsolateral side of their thorax with a 0.1 mm Minutien pin (Fine Scientific Tools, USA) dipped in the bacterial suspension. Sham-infections (injury controls) are carried out in the same fashion, except by dipping the pins in sterile MgSO₄ (10 mM) buffer. Across three experiments reported in this paper, five pathogens in total were used in this study, namely,

(a)   Enterococcus faecalis (Lazzaro et al., 2006), incubation temperature 37 ^OC;

(b)   Erwinia carotovora carotovora, strain Ecc15 (Martins et al., 2013), incubation temperature 29 ^OC;

(c)   Providencia rettgeri (Short and Lazzaro 2010), incubation temperature 37 ^OC;

(d)   Pseudomonas entomophila, strain L48 (Vodovar et al., 2005, Mullet et al., 2012), incubation temperature 27 ^OC; and,

(e)   Staphylococcus succinus, strain PK-1 (Singh et al., 2016), incubation temperature 37 ^OC.

These pathogens are reported natural pathogens of D. melanogaster and previous studies have demonstrated that these cause lethal infection in flies in the laboratory following systemic infection (Dionne and Schneider 2008, Troha and Buchon 2019). Some of the strains used here were isolated from wild-caught flies. Additionally, these pathogens all have intermediate virulence: they are neither completely benign nor completely lethal to the host. This is important since excessive lethality due to infection, or a lack thereof, can mask the subtle effects of starvation on post-infection survival. Furthermore, two of the pathogens used – E. faecalis and S. succinus – are Gram-positive bacteria, while the rest – P. entomophila, E. c. carotovora, and P. rettgeri – are Gram-negative bacteria. Since bacteria of different Gram-character elicit different immune responses from the insect hosts, with some overlap and crosstalk (Lemaitre and Hoffman 2007, Dionne and Schneider 2008, Vallet-Gely et al., 2008, Buchon et al., 2014), we included bacterial pathogens of both Gram-character in our experiments.

Previous studies have shown that in D. melanogaster, defense against E. faecalis remains unaffected by brief period of starvation prior to infection (Brown et al., 2009) and dietary restriction (Ayres and Schneider 2009), while brief starvation before infection improves survival and reduces systemic pathogen loads for E. c. carotovora (Brown et al., 2009). Additionally, low-protein diets increase susceptibility of flies to P. entomophila (Kutzer et al., 2018) and high-carbohydrate diets increase systemic pathogen loads following infection with P. rettgeri (Unckless et al., 2015).

4. Systemic bacterial load estimation

To measure the systemic bacterial load, females are first surface sterilized using 70% ethanol for 1 minute and 30 seconds, twice. Females are then washed in sterile distilled water for 30 seconds and dried using autoclaved tissue paper. Females are then transferred individually to 1.5 ml vials (microcentrifuge tubes) containing 50 or 75 microliters (depending upon pathogen used for infection) of sterile MgSO₄ (10 mM) buffer. Females are homogenized individually in these vials using a motorized pestle for 50-60 seconds. This homogenate is serially diluted (1:10 dilutions) 8 times in sterile MgSO₄ (10 mM) buffer. 10 microliters from each dilution, and the original homogenate, are spotted onto a lysogeny agar plate (2% agar, Luria Bertani Broth, Miler, HiMedia). The plates are incubated at required temperature for 8-12 hours (depending upon pathogen used for infection), and the number of colony forming units (CFUs) in each dilution is counted. The number of CFUs in the countable dilution (30 ≤ CFUs ≥ 300) is multiplied by appropriate dilution factor to obtain the bacterial load for each individual female.

5. Experiment 1: Effect of starvation and sexual activity on post-infection survival of females

Unmated females and males were obtained following the protocol described above. On day 12 PEL, half of the females were randomly assigned to ‘unmated’ treatment and the rest to ‘mated’ treatment. Females in the ‘mated’ treatment were combined with males in fresh food vials (1-2 ml standard food medium) in groups of 8 females and 10 males per vial, and allowed to mate for 4 hours (it was visually confirmed that each female had mated at least once). Following this, the females were lightly anesthetized and infected with bacterial pathogens (or sham-infected) following the infection protocol described above; males were discarded. Females from the ‘unmated’ treatment were similarly infected (or sham-infected). Following infections, half of the females from both these treatments were housed in vials with 1-2 ml of standard food medium (‘fed’ treatment), and the remaining were housed in vials with 1-2 ml 2% non-nutritive agar gel (‘starved’ treatment). This produced four experimental treatments:

(a)   Unmated, Fed (UF): 10 vials of infected and 5 vials of sham-infected females, each vial with 8 females;

(b)   Unmated, Starved (US): 10 vials of infected and 5 vials of sham-infected females, each vial with 8 females;

(c)   Mated, Fed (MF): 10 vials of infected and 5 vials of sham-infected females, each vial with 8 females; and,

(d)   Mated, Starved (MS): 10 vials of infected and 5 vials of sham-infected females, each vial with 8 females.

Note that in this experiment females were subjected to starvation from the time of infection. The vials were monitored for mortality every 4-6 hours, for 96 hours post-infection (HPI); alive flies were shifted to fresh food/agar vials at 48 HPI. This experiment was carried out for five bacterial pathogens – E. faecalis, E. c. carotovora, P. rettgeri, P. entomophila, and S. succinus – and replicated thrice for each pathogen. In each replicate, 320 females were subjected to infection (80 females x 4 treatments) and 160 females were subjected to sham-infection (80 females x 4 treatments). Uninfected controls (no infection, no sham-infection) were not included in the experimental design since preliminary experiments showed that uninfected females and sham-infected females do not differ in terms of survival.

6. Experiment 2: Effect of starvation and sexual activity on systemic bacterial load in infected females

Following a protocol identical to that of Experiment 1, females were distributed into four treatments (UF, US, MF, and MS, as described above) and infected; 100 infected females in each treatment and 30 sham-infected females in each treatment. Following infections, females were housed in plexiglass cages (14 cm length × 16 cm width × 13 cm height), with all females from a particular treatment in a single cage; infected and sham-infected females were housed in separate cages. (Cages of ‘fed’ treatments were supplied with standard food medium in 60 mm Petri plates and cages of ‘starved’ treatments were supplied with 2% non-nutritive agar gel in 60 mm Petri plates.) At 4 and 10 HPI, 12 infected females were randomly aspirated out of cages for each treatment and the systemic bacterial load was measured for individual females following the CFU enumeration protocol described above. Additionally, at 10 HPI, 6 sham-infected females were randomly aspirated out of cages for each treatment and the systemic bacterial load was measured for individual females. This experiment was carried out for three bacterial pathogens – E. faecalis, P. rettgeri, and P. entomophila – and replicated thrice for each pathogen. In each replicate, systemic bacterial load was measured for 96 individual infected females (12 females × 2 time-points × 4 treatments) and 24 individual sham-infected females (6 females × 1 time-point × 4 treatments). Sham-infected females from any treatment did not yield any CFUs during any of the experimental runs (3 pathogens × 3 replicates).